GNVC
>futility analyses at the interim looks?<
Yes, Dew, with this look the DSMB recommended the PACT trial continue.
Though I read you guys regularly, it's been awhile since I've posted on Biotech Values. Mostly, I haven't invested in companies that have been discussed here. Boy, I wish I had avoided this one! Today, I've come here to read a little hard analysis on GNVC. Everywhere else is full of company apologists in denial, IMO. I know DD isn't shy with his opinions and PGS used to know more about GNVC than anyone. (Sold over $4? Smart.) I unloaded half my position first thing this AM. I am now trying to give management's song and dance a fair consideration before unloading the other half.
I'd be grateful for any help. In particular, if anyone could help me understand now how the hazard ratio, 0.75, should be interpreted, that'd be great. I know the statistical definition. I just don't know how the FDA uses it.
This may be of help. Pardon me if it's beside the point. But just in case:
1) GNVC first looked at their data for PACT (a study of a TNF producing adenovirus injected percutaneously or through an endoscope for locally advanced, stage 3 pancreas cancer) in Dec '06 and (with a more complete presentation at ASCO '07 of the same look) declared SOC median survival 11 mos and active MS 19 mos with separation CI 75%. There were 52 subjects in the trial and only 12 events.
2) In Jan '08, the PACT trial had an FDA approved, no SPA, endpoint change from 20% improvement in 12 mo survival to overall survival p<0.25 minus penalties. The N stayed the same at 330 subjects with a 2:1 active/SOC enrollment. Looks at 92 and 184 events were built in with penalties of .001 and .004 respectively.
3) CEO informed investors sometime this year the trial was powered to show significance, with 300 events, at about a four month difference in median survival. Many investors had been hoping the first look would come close to that number.
4) The first look data discussed today (92 events, about 180 randomized) showed no difference in median survival (9.9 mos for both) and no difference in survival (about 10%) at 24mos (about three total survivors now). Although no K-M curve was shown, so we don't know specific data points, management described a "dramatic separation" of the curve between 12 and 18 mos that was responsible for this hazard ratio.
I understand PGS's point about the MOA not making sense when the drug, TNFerade, is given entirely in the first two months and the effect isn't seen until after ten months. Some might contend, weakly, that those who survived the rough active treatment had less disease and kept it at bay longer. And management flat out said they were reluctant to put out the 24 month number because there were so few subjects in the trial from two years ago. They strongly implied that soon those doing well at 18 months now will get to 24 months and help make that data point more significant (and likely show separation).
What I really want to understand is the strength of the hazard ratio number. If, somehow, PACT finishes with numbers similar to today's, is the FDA likely to consider GNVC's application favorably? Can the endpoint "overall survival with a P<0.245" be met when there is no difference in median survival? If so, how strong does hazard ratio have to be? I simply don't understand this number.
GNVC through out some examples. For instance, apparently Tarceva's HR was 0.81 in its trial for stage 3&4 pancreas cancer. I imagine, though, a drug already on the market has some advantages in getting approved for a third or fourth indication.
Lastly, GNVC's market cap is now a little less than twice cash. They claim the cash, about 25million, will last another 7-10 months. It's probably less. They NEED to partner very soon. They have very early, very good, but uncontrolled results in esophageal, H&N, rectal, melanoma and sarcoma. I'm sure they will offer all indications to any takers. They claim lots of interest. If they DO partner, the stock should do very well in the near term. Their vaccine side has some interesting projects...
I am also aware this company has it's back to the wall and will try very hard to keep this trial going, whether it's "wasted money" or not. I want out now if I decide no company with cash will likely be interested in partnering. I'm just not sure how anyone will consider that hazard ratio.
If you've read this far, thank you very much.
John
Market Data 
Markets 
AI
